Gorkon going to the Valley of the Sun

Yes indeed.  I will be flying out to Phoenix, AZ sunday for my next training class.  I’ll be learning IBM’s HACMP.  HACMP stands for High Availability Cluster Multi Processing.  Basically what it is is a way to fail over your server from one server to either a idle one (the proper way and also the most expensive way) or another LPAR  or running system.  It’s a way to fight hardware failure and also to ease maintenance on IBM’s pSeries machines.  For example, say I have a HACMP cluster set up between 2 servers, one running the application, the other running the database.  Say a firmware update comes out and I need to apply it before upgrading to a new Technology Level on AIX.  I can fail over the application to the database server, apply the firmware, bring the cluster back to normal and then do the same thing to the database and ideally this will happen and be transparant to the users.   Cool stuff….IF the application supports it.  Some do not.  The one I work on does not, BUT you can still cut downtime in that the application can be brought up in a reduced state on a server while you work on the other machine.

Oh and in case you did not know, IBM’s pSeries, then called RS/6000 was used to create the chess  playing program that beat Gary Kasparov.  Deep Blue as the system was called was based on a RS/6000 running AIX in a SP cluster.  SP stands for Super Parallel and is also the basis for super computers built on the same platform.   While I will probably never get to  play with a SP setup, it’s cool that I have 99 percent of the hardware I would need to do it.  You COULD do it over existing NICs, but I think it would be better to run Myrinet or some other low latency networking technology like infiniband instead of plain ole Gig copper.  Anyways, it’s not like I have a need to simulate nuclear explosions or do DNA research or research where I can find oil….many uses for super parallel computers.

Anyway, I should be podcasting on digital distraction and also blogging from Phoenix starting Sunday evening.

FLOSS Weekly with Jon “Maddog” Hall

If you have not already listened to the latest FLOSS Weekly over on the Twit network, do yourself a favor and go do that.  Jon Hall is a great hacker from the days of DECUS and was one of the first to work with Linux.  He has plenty of intresting stories and I listened with rapt intrest.  He has funny stories about Linus and some interesting history of how the sharing of software started way back when.

Also, since Chris and his wife just had a new child, Jon has offer Leo to help out with the FLOSS weekly which if you haven’t noticed turned out to be more like FLOSS monthly.  So hopefully we can get back to the weekly goodness that is the FLOSS Weekly.  Check the Jon “Maddog” Hall show out at:

http://tinyurl.com/366rga

Subscribe to  FLOSS at:

 http://tinyurl.com/2l3cxu